ANCUT1, a Fungal Cutinase MgCl₂-Activated by a Non-Essential Activation Mechanism for Poly(ethylene terephthalate) Hydrolysis

Abstract

Plastic waste, particularly poly(ethylene terephthalate) (PET), negatively impacts the environment and human health. Biotechnology could become an alternative to managing PET waste if enzymes ensure the recovery of terephthalic acid with efficiencies comparable to those of chemical treatments. Recent research has highlighted the potential of fungal cutinases, such as wild-type ANCUT1 (ANCUT1^wt) from Aspergillus nidulans, in achieving PET depolymerization. Fungal cutinases’ structures differ from those of bacterial cutinases, while their PET depolymerization mechanism has not been well studied. Here, a reliable model of the ANCUT1^wt was obtained using AlphaFold 2.0. Computational chemistry revealed potential cation-binding sites, which had not been described regarding enzymatic activation in fungal cutinases. Moreover, it allowed the prediction of residues with the ability to interact with a PET trimer that were mutation candidates to engineer the substrate binding cleft, seeking enhancements of PET hydrolysis. Enzyme kinetics revealed that both ANCUT1^wt and ANCUT1^N73V/L171Q (DM) were activated by MgCl₂, increasing the dissociation constant of the substrate and maximal reaction rate. We found that in the presence of MgCl₂, DM hydrolyzed different PET samples and released 9.1-fold more products than ANCUT1^wt. Scanning Electron Microscopy revealed a different hydrolysis mode of these enzymes, influenced by the polymer’s crystallinity and structure.

1. Introduction

Poly(ethylene terephthalate) (PET) is a semicrystalline thermoplastic composed of ethylene glycol (EG) and terephthalic acid (TPA) linked by an ester bond. It is a widely used plastic in various products due to its mechanical properties, transparency, and endurance [1]. However, due to how PET is used, particularly in single-use products, its accumulation in the natural ecosystems is causing significant environmental and human health issues [2,3]. This anthropogenic polymer has an estimated half-life of around 2500 years under natural environmental conditions [4]. Therefore, some treatment is required to reduce its disposal and prevent accumulation. Thermomechanical recycling and chemical hydrolysis could be a solution, even if continuous polymer recycling results in quality decay of the final product [5]. Chemical depolymerization requires high temperatures, high pressure, and toxic chemicals; generates by-products; and involves high water and energy consumption [4]. Many recent studies have focused on PET’s biological and enzymatic degradation and have developed standards for activity assessment and evaluation of the industrial scalability potential. The French biotechnology company CARBIOS exemplifies the optimistic perspective [6].

The biotechnological degradation of PET uses microorganisms, either bacteria or fungi identified as PET degraders, and enzymes attacking the polymer surface, where the obtention of monomeric units and their mineralization are the ultimate goals [2]. Cutinases, PETases, MHETases, and engineered PET hydrolases have been identified and studied [7]. These enzymes have the potential to release monomers through PET hydrolysis, and the approach could support repeated polymer recycling without loss of functional properties or added value, provided the monomers can be purified efficiently. While enzymatic PET degradation could offer a potential solution for the efficient management of polyester waste, there are still bottlenecks to improve the efficiency and yield of the process compared with chemical degradation [1]. Among the factors known to influence PET degradation and industrial PET recycling are those related to substrate properties such as microstructure, which can be characterized by the content of mobile amorphous PET fraction (Χ_MAF), rigid amorphous PET fraction (Χ_RAF), and crystalline PET fraction (Χ_C) [8], as well as that of biaxially oriented PET chains [9]. These parameters lead to differences in the physicochemical properties of PET and oligomers that will modify reaction phenomena, such as PET crystallization kinetics and non-enzymatic hydrolysis [6]. Furthermore, operations such as pretreatment by extrusion or micronization may increase the amorphous fraction or the exchange surface of PET [10]. Enzyme properties are fundamental to achieving success. The list is long and includes parameters such as catalytic efficiency for PET [10] and release of mono(2-hydroxyethyl) terephthalate (MHET), which may cause inhibition [11], thermostability [12], tolerance to high EG concentration [13], pH stability [6], specificity towards gauche and trans conformations of EG [14], Si- and Re-face nucleophilic attacks [15], electrostatic potential on the enzyme surface [16], and PET binding affinity [17]. These properties can be modified by enzyme engineering, directed evolution, and enzyme production to ensure the supply of enzymes for the PET recycling process at an industrial scale [10]. It is noteworthy that process parameters in PET depolymerization can affect the enzyme or enzyme preparations and products. These include medium composition, preferably water [6], reactor type [18], mass transfer [18], shear strength resistance [18], and pH control [13]. From an economic perspective, the enzyme–substrate ratio should be with fed-batch additions of high loads of PET with reuse of the reaction medium several times to enhance product recovery and purification [13]. The process may be designed to use an enzyme blend of PET exo- and endolytic activities [19]. To reduce costs, unit operations for enzyme recovery, enzyme immobilization, or nanoreactor design for catalyst reuse must be considered [18].

Despite being an environmentally friendly solution, the enzymatic PET degradation has not reached the cost-efficiency level required to make it the method of choice, and the search for new enzymes from natural sources, or by protein engineering, is still ongoing [6]. The discovery and study of different biocatalysts may improve our understanding of those challenging obstacles that prevent cost-effective PET degradation. Research has generally focused on actinomycete cutinases and Piscinibacter sakaiensis PETase, with few reports on fungal cutinases in PET degradation [2,7], which showed improvements in hydrolytic activity [20], thermal stability [21], substrate interactions [22,23,24], reaction conditions [25], enzyme kinetics [24], and molecular mechanisms of PET degradation modeled by quantum mechanics/molecular mechanics calculations [15]. Therefore, the present research aimed to investigate the fungal wild-type ANCUT1 cutinase (ANCUT1^wt) from Aspergillus nidulans to enhance PET hydrolytic activity. ANCUT1^wt [26] is an enzyme able to hydrolyze cutin [27], poly (butylene succinate-co-adipate) [28], poly (L-lactic acid), and poly (D-, L-lactic acid) [29], probably with a promiscuous polyesterase activity. Thus, through an in-silico study, the putative active-site residues contacting a PET trimer (3PET) were predicted, and ANCUT1^wt-MgCl₂ interactions improved the 3PET binding mode, and computational chemistry revealed potential cation-binding sites. We produced and tested an ANCUT1^N73V/L171Q (DM) cutinase variant through site-directed mutagenesis. Enzyme kinetics revealed that both enzymes were activated by MgCl₂, increasing the dissociation constant of the substrate (K_S) and maximal reaction rate (V_MAX). Therefore, magnesium ions act as a V/K-type activator. Our data showed positive cooperativity, with a Hill coefficient bigger than one, which suggests that both enzymes possibly bind more than one atom of Mg to the fully activated enzyme. The interaction between cutinases and MgCl₂ might be an allokairy-like behavior, so these cation-binding sites regulate enzymatic activity towards various substrates. These results indicate that activity improved in DM by non-essential cation activation, enhanced PET hydrolysis, and TPA and MHET production. Scanning Electron Microscopy (SEM) revealed a different hydrolysis mode of these enzymes, influenced by the polymer’s crystallinity and structure.

2. Results and Discussion

2.1. In Silico Analysis of ANCUT1^wt Interaction with a PET Insoluble Oligomer

The three-dimensional model of ANCUT1^wt was predicted using the artificial intelligence software Alphafold 2.0 [30] and validated with a per-residue confidence metric called the predicted local distance difference test (pLDDT) (pLDDT = 94.9). The model was validated and analyzed as shown in the Supplementary Information (Figures S1 and S2). The predicted three-dimensional structure of the mature ANCUT1^wt enzyme showed the characteristic α/β folding pattern of fungal cutinases, with a core of 5 β sheets surrounded by 8 α-helices. Compared to its homolog, the cutinase AfC from Aspergillus fumigatiaffinis (75% identity) [21], with Protein Data Bank (PDB) code (PDB ID: 8JCT), the putative catalytic triad comprises S109, H177, and D164. The S109 belongs to the conserved pentapeptide signature of fungal cutinases—GYSQG—lying between the β3 and α4. In the predicted ANCUT1^wt model, the Oγ of S109 is found at 3.02 Å from the Nε of H177, and the Oδ of D164 is at 2.94 Å from the Nδ of H177. This conformation closely resembles known cutinase and PETase active sites [31,32]. The amino acid residues at the predicted oxyanion cavity were identified as Q110 and S31; theoretically, they can stabilize the tetrahedral catalytic transition state. Three disulfide bridges were observed: C20–C98, connecting the loops 1–22 of the N-terminal with the one at 98–101 between α3 and β3 (where the S109 pentapeptide is located). A second bridge, formed between C46 and C59, stabilizes α1 and β2. This bridge is characteristic of the Aspergillus cutinases and is also present in the protein from Neosartorya fischeri [32]. According to mutagenesis experiments reported by Lee et al. [21], the second bridge impacts the thermostability and catalysis of AfC. Finally, the third bridge, between C160 and C169, is in the loop 160–175 between β5 and α7 and confers stability to D164 of the catalytic triad [33]. The cavity hosting the putative active site comprises mainly hydrophobic amino acid residues and is covered by L70, N73, F74, L171, and I173. Something similar occurs in the cutinase FsC from Fusarium solani with the amino acid residues L81, N84, L182, and V184, and this has been called a recognition site “mini-lid” [33]. This feature is also found in AfC [21], where residues N69, F70, and L167 are equivalent to N73, F74, and L171 in ANCUT1^wt. The mini-lid may restrict access to the ligand binding site, but the catalytic serine is still exposed to the solvent. We initially looked for amino acid residues on the ANCUT1^wt’s binding cleft model predicted by Alphafold 2.0 [30] able to interact with an insoluble trimer (3PET) [19] to improve hydrolytic activity against PET. They were predicted by in silico molecular docking using the Autodock VINA software version 1.1.2 [34]. A pose with Autodock-VINA’s binding energy of -6.4 kcal mol⁻¹ was chosen, given its interactions and binding mode, and considering the distance of the ester bond carbonyl carbon and the oxygen atom of the S109 side chain (3.48 Å). The geometry of the complex was optimized with the semiempirical PM7 (SQM) of MOPAC 2016 [35], and the enthalpy of formation of the complex, from its free species, was ΔH_fC = −479.86 ± 46.25 kJ mol⁻¹. This level of theory considers quantum effects, electronic effects, steric factors, and electrostatics. SQM geometry optimization is more likely to reproduce a native pose of the ligand in the enzyme [36].

Molecular dynamics (MD) simulated the ANCUT1^wt-3PET complex and the free enzyme using the Gromacs 2020.7 software [37]. The system was constructed as indicated below with NaCl (~150 mM) at 40 °C. During the simulation, we did not observe any significant drift from the starting ANCUT1^wt-3PET complex or the free enzyme, respectively.

The enzyme’s cleft cover remained stable for the time explored in the simulation. Then et al. [38], using the cutinase TfCut2 from Thermobifida fusca, and Oda et al. [39], using the cutinase CUT190* from Saccharomonospora viridis, found potential binding sites for Na⁺, Mg²⁺, and Ca²⁺. These groups reported some influence of ion binding on catalysis and thermostability. CUT190* transitions from a closed to an open state compatible with catalysis, and the holoenzyme hydrolyzed PET more efficiently [39]. Here, molecular dynamics simulations (MDs) were performed at 60 °C (close to the PET’s glass transition temperature (Tg) [40]) and 25, 150, or 300 mM MgCl₂, or without ions. The Mg²⁺ binds to four potential binding sites, as determined from residue contact (closer than 4.5 Å) frequencies [41]. Site I is formed by the amino acid residues D182 and D186, site II by S179 and D183, site III by the carbonyl of the G146 backbone, and site IV by Q77; in all sites, water molecules could stabilize the union according to the MD as shown in Figure 1. Moreover, site III is conserved in fungal cutinases reported to hydrolyze PET (PDB ID: 5X88, 4OYY, 3GBS, 1CEX, 8JCT, 3DEA, 5AJH), as shown in Figure S3, and the other ones are just present in AfC [21]. The new version, AlphaFold 3 [42], using the AlphaFold server, predicted an interaction between amino acid residues D183 and D186 of ANCUT1^wt and the Mg²⁺ ion with a pLDDT = 79.9, as shown in Figure S4, which means low confidence in the prediction. This last prediction combined sites I and II; AlphaFold 3 has made progress in predicting interactions between proteins and ions. The crystal structure complex (PDB ID: 7CY3) from fungal cutinase PCLE from Paraphoma sp. B47-9 shows a Na-binding site formed by residues S176, D179, and E180 without any reported functional properties. Those residues are equivalent to S179, D182, and D183 in ANCUT1^wt.

Sites I–III are close to the catalytic triad’s residue D164 and the C160–C167 disulfide bridge. Therefore, sites I-III appear as potentially important contributors to thermostability, which agrees with reports for TfCut2 [12]. Recently, Lee et al. [21] introduced the mutation of N182E to form a salt bridge with K153 in the AfC, which is present in the cutinase HiC from Humicola insolens. As a result, AfC increased its thermostability. From these observations, the last salt bridge and the ion binding site I might be candidates for replacement with a disulfide bridge. In addition, MDs with MgCl₂ 300 mM showed a closed-to-open conformation transition in ANCUT1^wt induced by Mg²⁺ ion binding, allowing a better fit of 3PET to the active cleft. Key residues to achieve this conformational change were in the mini-lid: L70, F74, N73, L171, and I173, which are oriented inside the binding site and block it, while in the open conformation, these residues create a wider space in the binding site. In addition to these residues, twenty additional residues are predicted to form the 3PET binding site, making them candidates for mutagenesis. The list includes hydrophobic residues (L75, P76, A113, V166, I173, and L178), polar neutral residues without charge (G30, T32, S39, T40, Q77, G78, T139, Q143, and G169), charged residues (E33, R140, K142, and E168), and finally, π-π stacking interactions appear to mediate F74 and Y109 residues with the aromatic rings of 3PET. Equivalent residues were identified in the interaction between bis(2-hydroxyethyl) terephthalate (BHET) and the FsC [24] and AfC [21] cutinases and, in (MHET)₂ contacts with HiC [43]. The sequence space available when all these residues are considered is vast, and an optimal biocatalyst for PET degradation is yet to be found.

MDs suggest Mg²⁺ as a possible activator, which has been confirmed experimentally vide infra. However, steric hindrance in the groove restrains a correct 3PET substrate binding. We hypothesized that a single swap between L171 and N73 would influence open conformation. Thus, a double mutant ANCUT1^N73V/L171Q (DM) was produced in silico to enhance substrate recognition. Val modifies the hydrophobic surface in this double mutant at the cleft’s mini-lid at α2 (close to the binding site). By contrast, the mini-lid’s contact segment (between β7 and α7) gains polarity, and Gln, with its longer arm, is more likely to form hydrogen bonds [44]. Additionally, the distance between the Cβ of residues L171 and N74 was predicted to increase from 5.65 Å to 5.75 Å by AlphaFold 2.0 [30]. MDs also revealed nine hydrogen bonds formed between the ANCUT1^wt amino acid residues and 3PET, with the highest frequencies being S31, E33, S39, T40, S109, T139, R140, K142, and Q143. Additionally, the distance from the S109 catalytic Oγ to the carbonyl carbon of the substrate fluctuates from 3.34 to 3.9 Å, within the range of distance values reported in the literature, going from 3.1 to 5.1 Å [31,32,33]. Furthermore, in MDs of the DM-3PET complex, new hydrogen bonds by S31 and Q110 were predicted to stabilize substrate binding. These form the oxyanion pocket, along with L70, V73, F74, and Q171. The distance from S109 to the ester carbonyl oscillates between 3 and 4.4 Å.

We show from the Root Mean Square Fluctuation (RMSF) analysis (Figure 2A,B) and the principal component analysis (PCA) of an ensemble of models from the trajectory that span principal component 1 for the ANCUT1^wt-3PET complex high mobility in residues 14–21 and 137–152 (Figure 2C, regions 1 and 4, red circles). At the same time, little flexibility was observed in the substrate recognition site and the catalytic triad. MgCl₂ induced higher flexibility into residues 14–21, 62–82, and 161–179 (Figure 2D, regions 1, 3, and 5, red circles) and stabilized region 4. With MgCl₂, a slight conformational change resulted in a better geometry in the 3PET binding to ANCUT1^wt (Figure 1C). In contrast, when we simulated the DM-3PET complex without MgCl₂, the ensemble structures from principal component 1 revealed more mobile regions were spotted at amino acid residues 14–21, 30–40, 62–81, 137–152, and 161–179 (Figure 2E, regions 1, 2, 3, 4, and 5, red circles). When the MgCl₂ interaction with the DM-3PET complex was simulated, we observed a reduced flexibility in amino acid residues in those same five regions as shown in principal component 1 (Figure 2F). The substrate binding mode for the DM-3PET complex with or without MgCl₂ appeared to be more stable, given its improvement in favorable contacts with the enzyme (Figure 1D). However, the ΔG_binding for 3PET was –355 ± 93 kJ mol⁻¹ for the ANCUT1^wt and dropped to –71 ± 52 kJ mol⁻¹ for the DM. This result indicates a significant loss of binding strength between the enzyme and 3PET in silico (Figure 2G). While this may seem inconvenient, it may not be so, because the lower enzyme–substrate stability should be accompanied by reduced enzyme–product stability, which may increase the catalytic rate constant (depending on the limiting catalytic step). Furthermore, a lower ΔG_binding is compatible with a broader specificity (i.e., a higher K_M), which may be advantageous for the degradation of mixed PET samples. Both factors may combine to increase the specificity constant k_CAT/K_M, which is technologically convenient. The single swapping between N73V and L171Q promotes a stable open conformation, and non-covalent interactions and a better ligand fit were observed by MDs.

These results led to the design and production of ANCUT1^N73V, ANCUT1^L171Q, and ANCUT1^N73V/L171Q variants, aiming at gaining insight into the mechanism of enzymatic PET hydrolysis.

Here, the Escherichia coli (DE3) pLysS strain was selected, allowing for the expression of recombinant ANCUT1^wt and its variants in the extracellular fraction, which were detected in the protein profile, zymogram, and Western blot (Figure S5). Soluble, functional biocatalysts were obtained, and the Western blot allowed the detection of the C-terminal His tag, subsequently used for purification by affinity chromatography.

2.2. Effect of MgCl₂ on Activity of ANCUT1^wt and the DM

Consistently, the non-essential activation analysis in our study indicated that both cutinases showed similar affinity, resulting in the dissociation constant of the substrate (K_S) from the DM with a value of 106 ± 29 μM, compared to ANCUT1^wt with K_S 191 ± 51 μM. No effect on V_MAX values was observed for the enzymes, V_MAX = 6393 ± 806 μM min⁻¹ for ANCUT1^wt and V_MAX = 6071 ± 643 μM min⁻¹ for DM, as shown in Figure 3. The introduction of the double mutation N73V/L171Q showed that the swapping produced an active enzyme with similar kinetic parameters to ANCUT1^wt for para-nitrophenyl butyrate (p-NPB) hydrolysis.

This global fit used for the non-essential activation kinetic model lets us calculate the activator dissociation constant (K_A), the change in K_S (α), and the change in V_MAX (β). The α values were 2.29 ± 0.76 and 3.17 ± 0.93 for ANCUT1^wt and DM, respectively. These α values suggest that binding Mg to the enzymes decreases the p-NPB binding affinity (K_S). The β values 2.82 ± 0.47 and 2.39 ± 0.29 were obtained for ANCUT1^wt and DM, respectively. These β values indicate that the enzymes increase the rate of the catalytic step shown as V_MAX. Hence, the ratio β/α is 1.0 for both enzymes, and the increase in activity offsets the decrease in substrate binding affinity. Therefore, magnesium ions act as V/K-type activators. Finally, the dissociation constant of the activator from the enzyme is K_A = 1.98 ± 0.75 mM for ANCUT1^wt. Albeit K_A was fixed to 0.65 mM for DM, it might be lower, suggesting an increase in activator binding affinity. As a result, this fixed K_A for DM results in no difference in the curves between 5 and 20 mM of MgCl₂. Further experiments are needed to evaluate a lower activator concentration to calculate K_A for DM. A somewhat unexpected activation effect of Mg²⁺ on the activity of both ANCUT1^wt and the DM variant, together with the finding of a higher relative affinity for Mg²⁺ of the DM by a factor of three, at least, suggests a non-essential cation activation.

The DM and the ANCUT1^wt do not differ in their activity against p-NFB. This result does not surprise us, because the introduced mutation should not prevent the entry of a soluble and small substrate such as p-NFB. However, a double mutation may enhance the hydrolysis of bigger molecules such as PET. After all, the bond to break is an ester anyway. Suzuki et al. [45] reported that CaCl₂ enhanced polyester hydrolysis with PCLE cutinase, while it lost its activity with MgCl₂, suggesting its ion-binding site may regulate polyester depolymerization. The non-essential cation activation usually entails allosteric binding of a metal ion, which likely induces conformational changes that enhance the enzyme’s activity beyond its basal state. This corresponds to what is shown in our fit, which uses non-linear regression against Equation (2), which considers the Hill model. Our data showed positive cooperativity with a Hill coefficient (n) bigger than one. These cutinases exhibited an allokairy-like behavior, characterized by a temporary transmission of a conformational change to the same site, or even to a different one [46], following the binding of MgCl₂, as shown in MDs. We observed two distinct patterns in Figure 3, suggesting that both enzymes may bind more than one atom of Mg to the fully activated enzyme, consistent with our MD observations. Since the concentrations of 25 and 30 mM MgCl₂ employed are relatively high, deviations to the Michalis-Menten equation were observed, the metal ion influences the solution’s ionic strength and, in turn, the buffer capacity. Therefore, a more detailed and in-depth kinetic analysis should attempt to dissect these contributions to reveal additional differences between the two enzymes. Further experiments in later work test the effect of the cation, ionic strength (e.g., tetramethylammonium or tetraethylammonium), and their interaction with pH and temperature on Mg activation. To our knowledge, no previous research has reported non-essential cation activation for fungal cutinases with allokairy-like behavior.

2.3. Effect of Reaction Temperature on Amorphous PET Hydrolysis

We initially studied the effect of a single mutation on commercial Goodfellow PET film (GF-PET); the hydrolysis was evaluated at 50 °C and 25 mM MgCl₂. ANCUT1^N73V and ANCUT1^L171Q showed 61.8% and 277% mixed phthalates (TPA, MHET, and oligomers), respectively, compared to the ANCUT1^wt. The N73V mutation had reduced enzymatic activity, suggesting a role for this residue in the catalysis. For instance, N84 in FsC, equivalent to N73 in ANCUT1^wt, was mutated to Ala, increasing 1.7 times its hydrolytic activity against PET [20]. Instead, the mutation N69A in AfC caused the activity against PET to decrease [21]. In ANCUT1^wt, the L171Q mutation increases its hydrolytic activity, pointing to a mutation hotspot at this residue in fungal cutinases. Leu mutated to Ala in FsC [20], enhancing PET hydrolysis, but the equivalent mutation in AfC (L167A) decreased PET hydrolysis [21]. This equivalent residue in AfC (L167) showed a large B-factor in the solved crystal (PDB ID: 8JCT), indicative of conformational flexibility when binding to PET [21], as we observed in MDs. Hence, we produced the double mutant ANCUT1^N73V/L171Q (DM), which showed stable open conformation, non-covalent interactions, and a tighter substrate binding in silico. Thus, the enzymatic hydrolysis of GF-PET film was evaluated as a function of reaction temperature. The DM exhibited higher GF-PET film hydrolysis in the temperature range from 40 to 65 °C with 25 mM MgCl₂ compared to ANCUT1^wt. The DM showed 9.1 times higher product release at 50 °C after 72 h of reaction, 320 ± 26 μM of mixed phthalates compared to 35 ± 2.5 μM of mixed phthalates for ANCUT1^wt (Figure 4). At 40 °C, the DM enzyme releases 4.1 times more phthalates than ANCUT1^wt. These results suggest that the open conformation has a positive effect on PET hydrolysis. We do not understand the mechanism by which MgCl₂ improves hydrolytic activity, so further experiments need to be performed, such as the obtention of a crystal complex with DM-substrate-MgCl₂ and ANCUT1^wt-substrate-MgCl₂. This may reveal more information to enhance the activity of fungal cutinases. At temperatures close to the Tg, there was a decrease in soluble mix phthalates due to their lack of thermal stability. As we observed in the p-NPB inactivation assay, the residual activity decreased drastically after 3 h of incubation. Therefore, engineering the protein to gain thermal stability could improve both the ANCUT1^wt and DM. We evaluated the effect of 0 mM MgCl₂ at 50 °C on PET hydrolysis; there was a decrease in product formation of around 50% for DM (150 ± 61 μM of mixed phthalates), although it released 4.3 times more phthalates than ANCUT1^wt (37 ± 9 μM of mixed phthalates).

The profile of soluble products at the end of the reaction analyzed by Reverse-Phase High-Performance Liquid Chromatography (RP-HPLC) showed 26 ± 2 μM of TPA by ANCUT1^wt and 237 ± 34 μM of TPA by DM, as shown in Figure S6. In the case of DM, an unidentified product was observed. We estimated the nature of the unidentified product from the ratio between the concentration determined by RP-HPLC and the one obtained by spectrophotometry. The ratio of 0.74 suggests a compound with roughly 26% phthalates. Schubert et al. [19] detected a soluble dimer and a trimer, which might correspond to the product of PET hydrolysis with DM.

Additionally, there are reports of non-enzymatic hydrolysis of BHET to MHET and of MHET to TPA due to the physicochemical conditions of the reaction [19]. Thus, 24-h reaction kinetics monitored by thin-layer chromatography (TLC) analysis and spectrophotometry method found products that were detected from the first hour of the reaction and up to 4 h, which might be the dimers and trimers reported by Schubert et al. [19] released by hydrolysis of internal bonds, but TLC did not detect that. These products could indicate that the enzyme has a lag phase for the release of products such as TPA, MHET, and BHET, as reported by Thomsen et al. [16].

By TLC analysis, we observed two products with similar spot intensity, with Rf values coinciding with those of TPA and MHET. The intensity of the products after 8 h remained constant. MHET is reported as the primary major product for other fungal cutinases [24,25] and other PET hydrolases [10,31], while other enzymes show low catalytic efficiency to hydrolyze MHET [47]. The reaction may have stopped due to product inhibition by MHET. Protein engineering has allowed the design of biocatalysts that improve catalytic efficiency to degrade MHET [11]. Furthermore, Sodium Dodecyl Sulfate Polyacrylamide Gel Electrophoresis (SDS-PAGE) analysis showed ANCUT1^wt and DM adsorbed to the PET surface.

As a control and comparison, we used the bacterial cutinase LCC^{F243I/D239C/S283C/Y127G} (LCC^ICCG), which has undergone engineering to enhance its thermal stability and activity, as reported by the biotechnology company CARBIOS [10]. Under the same reaction conditions at 50 °C with 25 mM MgCl₂, the enzyme LCC^ICCG released 5700 ± 530 μM of mixed phthalates from GF-PET film, with MHET being the primary product, according to TLC analysis, and without MgCl₂, it released 10810 ± 420 μM of mixed phthalates. As shown in Figure 4, MgCl₂ is a negative effector for this enzyme in GF-PET film hydrolysis. Lee et al. [21] reported that fungal-engineered AfC cutinase completely depolymerized PET film after 12 days at 60 °C, and Ronkvist et al. [9] reported that commercial fungal-engineered HiC cutinase completely depolymerized PET film after 4 days at 70 °C. There is still room to improve fungal cutinase activity against PET.

Figure 5 shows a hotspot map reported for fungal cutinases and known to enhance enzymatic activity. These residues have been related to greater accessibility to the binding site cleft and better substrate geometry for the catalytic Ser, formation of more or better hydrogen bonds with the substrate, and more stable hydrophobic, π-π, and/or π-CH₂ interactions with the substrate [20,21,44,48,49]. To improve the stability and thermostability of fungal cutinases, enhancing interactions in the hydrophobic core, the formation of salt bridges, disulfide bridges, the addition of prolines in loops, and the addition of divalent ions have been reported [21,26,38,50], with most mutations showing an additive effect. Figure S3 shows multiple amino acid sequence alignments of cutinases from Aspergillus nidulans [27] and another fungal cutinase, presenting crucial mutation points.

We used calorimetry to analyze the physical properties of the residual GF-PET after the enzymatic reaction. The Differential Scanning Calorimetry (DSC) thermograms recorded during the heating measurement of the GF-PET samples are shown in Figure S7. However, after enzymatic treatment with ANCUT1^wt and DM, there is no significant change in crystallinity compared to the one observed on the untreated GF-PET film, as shown in Table S5.

SEM analyzed the surface characteristics of GF-PET films after the enzymatic reaction. Changes were observed on the surface of the PET film. In the case of ANCUT1^wt, a smooth surface with some erosion and cracks was observed, as occurred with the use of the enzyme LCC^ICCG [16]. Surprisingly, enzymatic treatment with DM caused the generation of craters. The enzyme unable to attack some zones. One cause could be the presence of patches of crystalline PET. However, an additional cause could be the exposed -COO⁻ groups preventing enzyme adsorption to the surface, as reported for the PHL7 hydrolase [16] (Figure 6 and Figure S8). The electrostatic potential was calculated by the adaptive Poisson–Boltzmann solver (APBS) web services [51] for ANCUT1^wt and DM at different pH values, as shown in Figure S8. We observed negative potential in the binding cleft from DM at pH 9, which decreased as the pH decreased. Thus, the results observed by SEM suggest a different mode of action of the enzymes and agree with the kinetics of soluble product release.

2.4. Effect of Substrate Crystallinity and Polymer Structure on Enzymatic PET Hydrolysis

The enzymatic hydrolysis of PET starts with surface adsorption, and then the enzyme must diffuse to place its active site correctly onto the ester bond. Unobstructed sites are thus crucial for the reaction to occur. As substrate crystallinity increases, more crystalline nuclei predominantly present an EG in trans conformation and tighter packing [52]. Consequently, crystallinity is an obstacle to enzymatic hydrolysis. Here, the effect of substrate crystallinity on the hydrolysis of biaxially oriented PET film (BA-PET) (Χ_C = 34.3%) was studied, and product release was low throughout the temperature range evaluated (40 to 65 °C) for both enzymes. The maximum product release for ANCUT1^wt was at 60 °C (26 ± 2.9 μM of mixed phthalates) with MgCl₂, as shown in Figure 7, and the ion was important, since the product released decreased without MgCl₂ for ANCUT1^wt. In addition, DM released 18.4 ± 8.9 μM of mixed phthalates, and LCC^ICCG released 88 ± 1 μM of mixed phthalates with MgCl₂ at 50 °C, but without MgCl₂ no product was detected for DM or LCC^ICCG either. Many hydrolases have low activity against crystalline PET [47].

Moreover, no changes were observed in the thermograms of BA-PET samples or the crystallinity of the samples. Thomsen et al. [40] reported that even for LCC^ICCG at a reaction temperature close to the Tg, no soluble product was detected when the crystallinity range was between 23 and 27%, indicating that this type of substrate is more recalcitrant to enzymatic hydrolysis. The orientation of the polymer chains and their high Χ_RAF and Χ_C content negatively affect hydrolysis. Furthermore, we applied these cutinases from A. nidulans to hydrolyze postconsumer PET bottles (PC-PET) (Χ_C = 34.7%), and a negligible amount of mixed phthalates was detected. Thus, when thermal pretreatment is performed on PC-PET to increase the Χ_MAF of the polymer, only DM broke the PET polymer and released 359 ± 7 μM of phthalates, as shown in Figure 8. Therefore, it is necessary to continue searching for biocatalysts to hydrolyze PET with high crystallinity. Enzymes must overcome the energy barrier presented by non-covalent interactions in the crystalline phase and improve the activation energy of PET hydrolases to depolymerize this microstructure of PET efficiently [53]. To date, the solution has been a pretreatment to amorphized PET with different percentages of crystallinity by micronization and extrusion [10]. However, from an economic point of view, it is not cost-effective [54].

Under the reaction conditions evaluated for PET hydrolysis with different degrees of crystallinity, the potential of these two fungal enzymes was observed, and we identified some critical points to improve the performance of the cutinase ANCUT1^wt produced by Aspergillus nidulans and its variant DM. As mentioned before, designing a more robust biocatalyst and improving reaction conditions will lead to greater efficiency in PET hydrolysis and increased TPA productivity.

3. Materials and Methods

3.1. Modeling the Ligand–Protein Complex

The amino acid sequence of the ANCUT1^wt mature peptide was retrieved from the UNIPROT database with the Q5B2C1 accession code. This ANCUT1^wt protein and the ANCUT1^N73V/L171Q variant (DM) were modeled by AlphaFold 2.0 [30] using the ColabFold implementation [55]. According to a report by Schubert et al., the ligand used was an insoluble PET trimer (3PET) [19]. The model was built using the Avogadro software version 1.2.0 [56] and minimized with the MOPAC 2016 software [35]. The 3PET molecule was docked into the active-site cleft of the enzymes using the AutoDock tools version 1.5.6 [57] and AutoDock Vina version 1.1.2 [34] programs. Relevant poses were selected by the proximity of the ligand to the putative active-site catalytic triad.

3.2. Molecular Dynamics Simulations (MDs)

MDs were performed on the ANCUT1^wt and the DM, both in free form and in complex with 3PET, using the Gromacs 2020.7 software [37], with the ff14SB-ILDN force field [58]. The system was placed in an octahedral periodic box at 1.2 nm away from the box walls and solvated with the TIP3P water model. Na⁺ or Mg²⁺ plus Cl⁻¹ ions were added to neutralize the system. The system was pre-equilibrated at constant temperature and standard pressure conditions (60 °C and 1 bar) using a V-rescaled thermostat and Berendsen barostat. MDs were run for at least 30 ns in 3 or more replicates using GPU hardware. The trajectory analyses were performed using the Gromacs 2020.7 root mean square deviation (RMSD), root mean square fluctuation (RMSF), and radius of gyration (Rg) tools. The entropy was calculated in the convergent phases of the conformers of free 3PET or complex enzyme-3PET and free enzyme residues with a 1.5 Å restraint using the clustering extension [59] implemented in VMD software version 1.9.3 [60]. Principal component analysis (PCA) was performed using GROMACS covar and GROMACS anaeig codes. Cα was the group chosen for the least square fit and the enzyme-3PET complex for the covariance analysis. The quasiharmonic approximation to entropy (Schlitter formula) was run in GROMACS 2020.7 software [37]. The hydrogen bonds, salt bridges, and contacts were analyzed as reported elsewhere [41] using VMD scripting capabilities [60].

3.3. Semiempirical Quantum Mechanical (SQM) Calculations of Enzyme–Ligand Models

The enthalpy of complex formation ∆H_fC was estimated using SQM with the localized molecular orbitals approach (LMO), the LMO-PM7 method, and the conductor-like screening implicit solvent model (COSMO) implicit solvation [35]. The enthalpy was obtained from the difference in the PM7 heat of formation of the complex (H_fC) minus that of the free enzyme (H_fE) and the isolated ligand (H_fL), given by Equation (1):

∆H_fC = H_fC − H_fE − H_fL

(1)

Before the computation, the geometry of the complex was optimized to a gradient of 10 kcal mol⁻¹ Å⁻¹, using the default algorithms.

3.4. Plasmid Construction and Mutant Enzymes by Site-Directed Mutagenesis

The gene encoding ANCUT1^wt (GenBank code: PP502430.1) was manually modified and optimized for expression in Escherichia coli (Figure S9). According to Tournier et al., the genes encoding the LCC^{F243I/D239C/S283C/Y127G} cutinase variant (LCC^ICCG) were also modified and optimized [10] and all were synthesized by IDT (Integrated DNA Technologies, Inc., Coralville, IA, USA). The genes of ANCUT1^wt and DM enzymes were cloned into E. coli (DE3) plysS, and the LCC^ICCG gene was cloned into E. coli (DE3) BL21, according to Tournier et al. [10].

The variants ANCUT1^N73V, ANCUT1^L171Q, and ANCUT1^N73V/L171Q were generated from the pET22b-ancut1 plasmid by PCR amplification in a Techne TC-312 thermocycler (Techne Inc, Burlington, NJ, USA). The incorporation of point mutations was carried out using mutagenic oligonucleotides (Table S1) following the protocol of the Q5 site-directed mutagenesis kit (New England Biolabs, Ipswich, MA, USA), modified for this study according to Tables S2–S4. The DNA plasmid integrity was analyzed by sequencing (Macrogen, SEL, ROK).

3.5. Expression and Purification of Enzymes

First, every strain was inoculated in 25 mL LB (tryptone 1%, yeast extract 0.5%, NaCl 1%) broth pH 7.2 in 250 mL flasks at 37 °C, 300 rpm (New Brunswick™ Innova^® 40R, Edison, NJ, USA), and 100 μg mL⁻¹ ampicillin. When optical density at 600 nm reached 0.6 A.U. (absorbance units), the cultures were incubated at 30 °C and induced with 0.4 mM isopropyl-β-D-1-thiogalactopyranoside (IPTG) for 21 h with orbital shaking at 300 rpm (New Brunswick™ Innova^® 40R, Edison, NJ, USA). Then, the supernatant was collected by centrifugation at 10,016× g for 15 min. Enzymatic activity was determined in the supernatant and cell pellet. We observed good activity in the supernatant for ANCUT1^wt and DM, which facilitated the purification process, and LCC^ICCG presented higher activity in the cytosol fraction. The supernatant with activity was concentrated by ultrafiltration under stirring (Amicon Ultrafiltration Cell and Amicon YM-10 NMWCO 10 kDa membrane, Millipore, Burlington, MA, USA). Finally, enzymes were purified through Protino Ni-TED chromatography (Macherey-Nagel, Düren, DE) in 3 mL bed columns. Briefly, the column was equilibrated with 12 mL of Lysis Equilibration Wash (LEW) Buffer (50 mM NaH₂PO₄, 300 mM NaCl, pH 8). After the sample application, the column was washed with 24 mL of LEW buffer and eluted with 12 mL of a 10 to 250 mM imidazole gradient in LEW buffer. The recombinant enzyme fraction was collected and concentrated by diafiltration and ultrafiltration, as already described.

Protein samples were analyzed by Sodium Dodecyl Sulfate Polyacrylamide Gel Electrophoresis (SDS-PAGE) using Coomassie staining. Zymography as reported elsewhere [26]. The enzymes were further analyzed by Western blot using anti-His tag antibody (mouse monoclonal antibody, alkaline phosphatase conjugate, Life Technologies, Carlsbad, CA, USA). The protein concentration was determined according to the Bradford method [61] using a Bio-Rad protein assay kit (Bio-Rad Laboratories, Hercules, CA, USA), with bovine serum albumin as a standard.

3.6. Enzyme Kinetics

The enzymatic activity was measured at 60 °C in 50 mM Tris-HCl buffer pH 9 and as a function of variable concentrations (10, 30, 50, 70, 100, 150, 200, 300, and 500 μM) of para-nitrophenyl butyrate (p-NPB) (Sigma Aldrich, STL, MO, USA) and 0, 5, 10, 15, 20, 25, or 30 mM of MgCl₂ (as indicated below). The enzymatic assay was conducted in multiwell microplates, with 170 μL of buffer with or without MgCl₂, 10 μL of the enzymatic solution, and 20 μL of p-NPB stock (dissolved in 100% ethanol) in each well. The rate of substrate non-enzymatic hydrolysis was evaluated in a buffer lacking the enzyme. Though small, this rate was subtracted from that of enzymatic assays. Each enzymatic assay was performed at least in triplicate, and the kinetics were monitored by absorbance at 420 nm every 3 s for at least 1 min using the Gen5 1.10 software to operate an Epoch spectrophotometer (Biotek, Winooski, VT, USA). One enzymatic activity unit (U) was defined as 1 μmol of p-NPB converted to p-Nitrophenol (p-NP) per minute under the reaction conditions. Kinetic parameters were calculated using non-linear regression calculations using GNUplot software version 5.4 [62]. The initial velocity (v₀) data were fitted to the non-essential activation kinetic model. In Equation (2), A represents the activator; S is a soluble substrate; K_S is the dissociation constant of the substrate from the enzyme; K_A is the dissociation constant of the activator from the enzyme; α and β represent maximal changes in K_S and V_MAX of the enzyme in the presence of different concentrations of MgCl₂; and finally, n is the Hill coefficient. The deduction of Equation (2) is shown in Supplementary Materials.

$$v_0={\displaystyle \frac{V_{MAX}\frac{\left[S\right]}{K_S}(1+\beta \frac{{\left[A\right]}^n}{\alpha {K_A}^n})}{1+\frac{\left[S\right]}{K_S}+\frac{{\left[A\right]}^n}{{K_A}^n}+\frac{{\left[A\right]}^n}{\alpha {K_A}^n}}}$$

(2)

3.7. Depolymerization of PET Films by ANCUT1^wt, DM, and LCC^ICCG

The following materials were tested to study the degradation of PET: 0.25 mm thick commercial PET film (Goodfellow Cambridge Ltd. (Huntingdon, UK, ES301445) (GF-PET), 0.05 mm thick biaxially oriented PET film (Goodfellow Cambridge Ltd., Huntingdon, UK, ES301250) (BA-PET), and postconsumer PET bottle (PC-PET) (Pepsico Inc., Epura, CDMX, MEX). The material was cut into 5 mm × 5 mm squares (approx.). The reaction mixture contained 50 mM Tris-HCl buffer, pH 9, 0.5 μM of enzyme, 0 or 25 mM MgCl₂, and 50 g L⁻¹ of PET sample in a reaction volume of 1 mL. The reaction was incubated from 40 to 65 °C for 72 h at 800 rpm in a Thermo Scientific Peltier heated shaking dry bath (ThermoFisher Scientific, Waltham, MA, USA). Reactions were assayed in triplicate and stopped by adding 1 mL methanol. The samples were centrifuged at 14,500× g for 10 min, and the soluble fraction was analyzed by photometric absorption at 241 nm, where mixed phthalates are terephthalic acid (TPA), mono(2-hydroxyethyl) terephthalate (MHET), and oligomers, detected, quantified, and identified by Reverse-Phase High-Performance Liquid Chromatography (RP-HPLC) and/or thin-layer chromatography (TLC) plates (silica gel 60 F₂₅₄ aluminum sheet, Merck, San Jose, CA, USA).

Quantification of soluble products was analyzed by RP-HPLC using a 1260 Infinity (Agilent Technologies, Santa Clara, CA, USA) equipped with a Chromolith RP-18e 50-4.6 mm column. Samples were injected in 10 μL volumes and eluted at 1 mL min⁻¹ and 25 °C, employing a mobile phase consisting of 0.2% aqueous acetic acid: methanol starting with 4:5 (20 min) and followed by a 1:10 mixture. Analytes were detected by a diode array optical unit at 244 nm. The TLC solvent system was chloroform/methanol (60:40 v/v).

PET pretreatment was as follows: A piece of an aluminum plate (~10 × 7 cm) was placed on a heat block at 290 °C, and 2 g of PC-PET were laid over the aluminum plate and allowed to melt (1–2 min). The sample was covered with a second aluminum plate pressed uniformly at 14.2 g cm⁻² for 1 to 2 min and heated again for 10 min. Finally, the plates were cooled by immediate transfer to an ice-water bath and manually shaken using metal tweezers for about 5 min. The appearance of a yellowish material is expected. When the material had a white–opaque appearance due to unwanted crystalline inclusions, it was discarded.

3.8. Differential Scanning Calorimetry (DSC)

The PET samples were analyzed by DSC using DSC1 Mettler Toledo equipment (CMH, USA). The analyses were performed in a 0 to 280 °C temperature range under a nitrogen atmosphere. All determinations were made with a heating rate of 10 °C min⁻¹. The mass of the films was 4.87 ± 0.69 mg. The melting enthalpy ΔHm, the crystallization enthalpy ΔH_c, and the areas under the curve of the endothermic and exothermic peaks (J g⁻¹) were considered to determine the degree of crystallinity. The Tg was calculated according to the equipment’s manufacturer. The crystallinity fraction was calculated using Equation (3).

$$crystallinity\left(\%\right)={\displaystyle \frac{\left({\Delta H}_m\right)-\left({\Delta H}_c\right)}{{\Delta H}_m^{°}}}\times 100$$

(3)

where ΔHm° is the melting enthalpy for 100% crystalline PET, using an estimated value of 140.1 J g⁻¹ [9]. The graph was made using the Origin software v8.

3.9. Scanning Electron Microscopy (SEM)

To analyze the morphology of the PET films before and after enzymatic treatment, they were washed with deionized water, 5% SDS, and then with 70% ethanol, allowing them to dry at room temperature for 24 h. The samples were examined by SEM in a JEOL JSM-5910 electron Microscope (Tokyo, Japan) with an acceleration voltage of 10 kV. The samples had been previously coated with a nanometric layer of gold under a high vacuum.

4. Conclusions

In this study, we employed a predicted 3D model using AlphaFold 2.0 and MD to identify 25 amino acid residues in the substrate binding cleft, with particular focus on residues L70, N73, F74, L171, and I173, which block the binding cleft of ANCUT1^wt. To overcome this limitation, we engineered a double mutant variant, ANCUT1^N73V/L171Q (DM), which enhances substrate binding. MDs revealed four potential binding sites and suggested that in the presence of Mg²⁺ an enzymatic activation could be elicited, a phenomenon that had not been previously reported in fungal cutinases. The binding causes a conformational change from a closed to an open binding cleft in ANCUT1^wt. This finding was corroborated by enzyme kinetics, which showed that magnesium ions (MgCl₂) activated both ANCUT1^wt and DM by increasing their K_S and V_MAX, consistent with the non-essential activation model, acting as V/K-type activators. A Hill coefficient bigger than one suggests that both enzymes possibly bind more than one atom of Mg to the fully activated enzyme. These cation-binding sites regulated enzymatic activity with allokairy-like behavior. PET hydrolysis assays revealed that the DM variant, activated by MgCl₂, is 9.1-fold more active than the ANCUT1^wt in hydrolyzing a GF-PET film at 50 °C, releasing TPA and MHET as the predominant products. ANCUT1^wt and DM hydrolyzed amorphous PET. SEM results showed distinct modes of action for these cutinases for PET hydrolysis, influenced by temperature, substrate crystallinity, and polymer structure. Notably, DM exhibits increased activity against various substrates, and the presence of MgCl₂, while not strictly necessary, often enhances the enzyme’s activity. Therefore, DM non-essential activated by MgCl₂ emerges as a promising biocatalyst for PET depolymerization, supported by a 3D model predicted by artificial intelligence and computational chemistry. We believe that these cation-binding sites and mutagenesis targets could enhance PET hydrolytic activity in fungal cutinases.